The world of big data analytics and data science can be fascinating but over-hyped in equal measure. But what if we could summarize the seven key 'cultural elements' of big data that express how firms should culturally curate and embrace the potential business advantages that are on offer.
Woolly & hackneyed notions
The problem with firms playing in the big data analytics space is the seemingly infectious predilection they have for talking about these (arguably) woolly and hackneyed notions they call business outcomes, so-called ‘business transformation’ and this thing known as the ‘customer journey’. Database and data services company Teradata pretty much has that problem too - you’d be worried (and possibly disappointed) if it didn’t - but there is grease beneath the gloss and real meaning if you scratch hard enough.
Hiding behind a series of product, service and licensing strategy announcements made recently, it is possible to triple-distill a seven point plan for looking not at big data databases and their analytics engines as such… but at the cultural trends that define where effective big data is happening in the workplace.
1 - Business outcomes & experimentation
Teradata president and CEO Victor Lund argues that firms who want to bring a big data analytics culture into the workplace need to think about it as a process of business experimentation, but one that has a notion of what might be possible on the road ahead.
"Firms need to have some notion of what business outcomes they want to get to in the first instance before they start to bring data analytics culture online," said Lund.
Even if those business goals change over time (as they inevitably will), firms need to have an idea what what new markets, new working methods and new efficiencies they think they want to achieve. Essentially it is a process of business experimentation and this is why Teradata says it has structured its service offerings to include a fully configured, ready to run developer version of the Teradata Database for evaluation and testing new data theories.
2 - Look for repeatable solutions
As confusing as it all looks, a lot of big data analytics has already been done. To be clearer, what we mean is that a lot of big data analytics processes have been applied by other businesses in similar use cases. When we talk about this word that keeps coming up in technology these days 'automation', this is part of what we mean here i.e. we can automate what happens to certain types of data for certain data workloads based upon predefined models that have been in use elsewhere. This is not Teradata 'physically sharing' its customer's data around with other customer's because exact data values are not required. Instead it is the so-called 'data model' inside which that data is analyzed.
“This is what the future of customer analytics looks like. For years, advanced analytics has been the elite domain of data scientists and programmers. Teradata has continued to focus on democratizing analytics,” said Robin Bloor, founder and chief analyst, Bloor Group. “Teradata is creating repeatable analytic solution templates which leverage intellectual property from previous implementations – with consulting services, program logic, schema, visualization and smart interfaces. This approach accelerates time to value.”
3 - Applications built around analytics
Teradata has suggested that, for success in this space, firms need to build their software applications around a central connection to data analytics itself. If this happens (and yes, it is a big if) then there is (arguably) a higher probability that the business function in any given organization will start to accept the importance of data analytcs in terms of a) feeding data into it and b) using the insight results it is supposed to offer.
“We want businesses to grow by delivering more sales, reducing churn and improving customer satisfaction. The latest release of our Customer Journey offering sees Teradata putting more analytics into the hands of marketing with easier access to analytics, dynamic visualizations, machine learning and predictive simulations,” said Dan Harrington, executive VP of consulting & support Services at Teradata. “Our solution brings together all the required technology, plus the consulting expertise to achieve faster time to market."
4 - Multiple data operating environments
To make big data analytics work effectively, analytics engines need to be able to span multiple data formats, work with multiple types of data storage (on-premises cloud, public cloud, hybrid mixtures -- and different data disk types too) and multiple data types (structured data, semi-structured data and messy hard to classify or digitally quantify unstructured data)... so, says Teradata, it's all about being able to work in multiple data operating environments.
5 - Cloud compatible licensing
For big data analytics to work, culturally, it has to be able to work anywhere. What this means in technical terms is that it has to work with the way firms are using cloud computing today i.e. a mixture of public, private and hybrid as mentioned in rule #4.
The difficulty here stems from the fact that different clouds will be supplied with different licensing models. This is why Teradata is making a big deal about its portable database license flexibility across hybrid cloud deployments.
6 - Thick data
Big data culture needs to embrace the idea of so-called thick data too.
This type of data is defined as qualitative information that provides insights into the everyday emotional lives of consumers. It seeks to go some considerable way beyond big data as it attempts to explain why consumers have certain preferences, the reasons they behave the way they do, why certain trends stick and so on.
"Big data has become so big that we have gotten to a point where we regard absolutely everything as potentially measurable -- and that is really dangerous," insists self-styled 'technology ethnographer' Tricia **bleep** in her capacity as founder of Constellate Data. "Computers have the power to scale, but we humans have the depth. Computers do not have the ability to provide 'explanatory information', so thick data allows us to embrace emerging indicators that have yet to be even quantified."
**bleep**'s point brings us full circle back to Teradata's proposition that we need to think about business outcomes in the first place.
"The trouble is, data scientists rarely sit down and talk to business people. If a technical book company wants to sell more books, then the business question it really should ask itself is: how do people want to learn today? If a car company wants to sell more cars, it should ask: how do people perceive mobility today and is being alone in a vehicle [for some quality 'me time'] also an important component element of the total automotive experience alongside the journey? It's important that [data scientists and business people] remember to get together and co-create," said **bleep**.